The NICHD-sponsored Multicenter Reproductive Medicine Network provide the opportunity to investigate definitively with sufficient power the role of progesterone deficiency (LPD) in the etiology of spontaneous abortion. In subjects having at least one spontaneous abortion we will further determine not only sensitivity and specificity for currently nonvalidated diagnostic tests but also efficacy of medical (progesterone) treatment. In this model, the investigators propose a prospective randomized study pooling the expertise of the Divisions of Reproductive Endocrinology and Reproductive Genetics, both in Department of Obstetrics and Gynecology at the University of Tennessee. This trial begins with recruitment of spontaneously aborting (first trimester) women whose abortus will be karyotyped by aggressive approaches (ultrasound-guided CVS or curettage designed to achieve 90% culture success rates). Prior to a next planned pregnancy, extensive medical and laboratory investigation will exclude known causes of repeated losses (infection, autoimmune factors, leiomyomas, uterine anomalies, and parental chromosomal rearrangements). In each of two cycles, four diagnostic methods currently used to evaluate presence or absence of luteal phase defect will be utilized: endometrial biopsy,, serum progesterone, luteal phase length, and endometrial receptors. Results of these tests are blinded until the next pregnancy outcome is known, a process necessitated by lack of a single "gold standard" for diagnosis of LPD. Confounding variables (e.g., maternal age, previous losses smoking, alcohol intake) will be recorded, and subjects then randomized to progesterone or placebo treatments. Subsequent pregnancies are then monitored weekly for hormonal and ultrasonographic status during the first trimester. Any spontaneous abortion will be subjected to chromosomal analysis, again by ultrasound- guided CVS for villus recovery prior to passage of tissue. The salient comparison will be the frequency of euploid losses in the placebo group vs. frequency of euploid losses in the progesterone-treated group. If LPD is a valid entity of clinical significance, the frequency in the latter should be lower. If LPD is associated with an increased number of euploid losses, we will next determine which of the 4 diagnostic methods is most sensitive and specific. Requiring entry of approximately 400 spontaneously aborting women and thus likely a multicenter design, this project will be the definitive study concerning success of failure of progesterone therapy in preventing euploid abortion. Our study will also show any existing association between aneuploidy and LPD, a relationship which would suggest LPD could be a secondary phenomenon in which defective embryos lead to corpus luteum dysfunction. The coexistence of Reproductive Endocrinology and Reproductive Genetics in our department will enable the network to incorporate immunogenetic and molecular genetic technology, not only to the current proposal, but also to others elected by the cooperative Reproductive Medicine Network.